Introduction

Women currently constitute approximately half of all gamers globally and 51 per cent of gamers in Canada. Despite this parity, gaming culture remains hostile, with 58 per cent of women gamers reporting annual experiences of toxicity—a figure that has remained unchanged for a decade. Research indicates that 40 per cent of women experience verbal abuse, 32 per cent receive inappropriate messages, 28 per cent suffer sexual harassment, and 15 per cent receive rape threats. Younger players aged 16–24 are disproportionately affected, with 49 per cent reporting regular toxicity. The persistence of these figures suggests that individual bad behaviour is less the cause than the technical infrastructure and platform affordances that structurally reproduce harm.

Theoretical framework: Affordances, design and gendered harm

The concept of affordances refers to the action possibilities and constraints that platform features invite or shape. Gendered harassment in gaming is not merely a behavioural issue but an affordance problem. Default design choices, such as default-on voice chat and anonymised matchmaking, create the conditions for harm. Rather than focusing on ex-post content removal, this framework examines how platform design enables harmful interactions ex-ante. Gaming culture often constructs a “hegemonic masculine identity” that treats women as intruders, a dynamic supported by specific technical environments that act as “misogyny incubators.”

How gaming affordances produce gendered harm

Gaming platforms employ distinct affordances that structure exposure to harassment, primarily through voice communication systems, identity and anonymity features, and algorithmic recommendations. These elements combine to make harassment easy to perpetrate and difficult for victims to avoid or for platforms to address effectively.

Voice communication: The primary vector

Voice chat is the most significant vector for gendered harassment. In studies of games such as Valorant and Overwatch, sexual comments were heard in 14.17 per cent of matches, and general verbal abuse occurred in 25.8 per cent. Identity-based harassment was found in approximately 50 per cent of hour-long sessions. Consequently, 61 per cent of LGBTQ+ gamers feel uncomfortable using voice chat. Design solutions exist, such as the in-game voice changers introduced in Marvel Rivals, but their lack of widespread adoption reflects a governance gap. Recommendations include adopting safety-by-design standards, such as opt-in rather than default-on voice identification and built-in voice-masking tools.

Anonymity, identity and coordinated harm

Anonymity reduces the social and reputational costs of harassment. Coordinated campaigns exploit the differing affordances of multiple platforms simultaneously, using Steam for organising, Discord for coordination, and 4chan or X for content generation and amplification. The 2024 campaign against Sweet Baby Inc. demonstrated this cross-platform logic, where a Steam curator group grew to over 355,000 subscribers, facilitating harassment that reached millions of users.

Algorithmic recommendation and discursive radicalisation

Engagement-optimised algorithms tend to route users toward extreme content to increase platform time and advertising revenue. Casual exposure to anti-diversity commentary can escalate into organised misogynist movements. Experimental research found that YouTube’s recommendation system routinely routed accounts interested in gaming toward progressively more misogynist and “incel” material via its Shorts feature. Platforms should be required to conduct affordance impact assessments to evaluate how these algorithms affect gendered participation and safety.

AI as affordance: New risks and governance gaps

Artificial intelligence (AI) is increasingly embedded in gaming as both a moderation tool and an interactive feature. These roles introduce new governance challenges that current frameworks have failed to address adequately.

AI moderation: Structural limitations

Roblox processes roughly six billion text chat messages and 1.1 million hours of voice communication daily, making AI moderation a structural necessity. However, current systems rely on “dictionaries of profanities” that miss contextual, ambient misogyny and low-level hostility. AI often makes oversimplified judgements, missing structural patterns of targeting. To combat this, platform operators should mandate transparency and independent audits for AI moderation, disclosing design parameters and training data to ensure equitable performance across identity categories.

Generative AI and the new misogyny infrastructure

Generative AI is rapidly entering the sector; by July 2025, 20 per cent of new releases on Steam disclosed the use of generative AI tools. Systems like NVIDIA’s Avatar Cloud Engine create dynamic relationships that adapt to player behaviour, risking the reproduction of misogynistic discourse. The scale of this risk was demonstrated in January 2026 when the Grok system on X created approximately three million sexualised images in eleven days, including 23,000 depicting minors. This suggests that governance mechanisms designed for human content are insufficient for the volume of AI-generated content.

The governance gap: Why content frameworks miss affordance harms

Existing frameworks focus on removing specific items of content rather than addressing the design choices that ensure harmful interactions recur. Canada’s proposed Online Harms Act (Bill C-63) lapsed in 2025 and was limited by its focus on overt material and child protection, leaving adult harassment and gaming-specific misogyny largely out of scope. It also lacked mechanisms to address AI moderation failures.

Comparative governance: Lessons on structural accountability

The EU Digital Services Act (DSA) and the UK Online Safety Act 2023 offer superior models. The DSA requires large platforms to mitigate “systemic risks,” including gender-based violence, with fines of up to six per cent of global annual revenue. The UK Act explicitly classifies gaming services as user-to-user services, allowing regulator Ofcom to impose fines of up to 10 per cent of worldwide revenue. A structural accountability framework should be established in other jurisdictions to classify gaming services as user-to-user platforms subject to risk assessment and duty-of-care obligations.

Conclusion

Gendered harm persists because governance has addressed the symptoms rather than the underlying platform affordances. Until design choices—like default voice chat and hostile recommendation algorithms—are subject to accountability, toxicity levels will likely remain static. The emergence of generative AI adds a new dimension to this challenge, requiring dedicated accountability mechanisms and meaningful consultation with affected communities, including women and LGBTQ+ gamers, as a foundational step in any design process.